Developed and evaluated a scientific-domain RAG pipeline to study citation grounding and faithfulness of LLM-generated answers using retrieved evidence.
Implemented retrieval and evaluation workflows involving sentence embeddings, vector indexing, automated QA generation, and NLI-based citation verification.
Performed empirical analysis of retrieval methods, reranking approaches, and model behavior to identify factors affecting trustworthy AI-generated responses.
Machine LearningPythonDeep LearningRAGNLP
Tayari24
Full Stack Developer
May 2025 - Jul 2025 (2 m) • bengaluru, karnataka, india